gmdh_example

  GMDH_main为GMDH主函数； variable_Combin为输入层初始变量选为x1,x1^2,x1*x2,x2^2,x2时的输入变量矩阵值 variable_select计算X_train训练输入数据,Y_train训练输出数据,X_test测试输入数据,Y_test测试输出数据 Combin为求变量的两两组合 Sym_Combin为求符号变量的两两组合 PE_AIC求每层各神经元的参数估计W，及训练数据在参数估计后估计的输出out_train，测试数据在参数估计后的估计输出out_test，还有与实际比较的误差平方和PESS， 以及准则值AIC sym_representation求最终的输入输出符号表达式 Criterion_value求准则值 (GMDH_main main function for the GMDH variable_Combin initial variables chosen for the input layer, x1, x1 ^ 2, x1* x2, x2 ^ 2, x2 the value of the input variable matrix calculation X_train variable_select training input data, Y_train training output data, X_test test input data, Y_test for the sake of the test output data Combin pairwise combination of variables for the sake of symbolic variables Sym_Combin pairwise combinations PE_AIC neurons find each parameter estimation of the W, and after the training data is estimated in the estimation of output out_train, test data After the parameter estimates of the output estimated out_test, also compared with the actual sum of squared errors PESS, as well as the final criterion value AIC sym_representation seek the input and output values ​ ​ of symbolic expressions Criterion_value find criteria)

传递函数为线性函数_特例验证
...........................\Combin.m,290,2011-03-27
...........................\Criterion_value.m,252,2011-03-28
...........................\GMDH_main.m,9357,2011-03-29
...........................\network_search.m,5555,2011-03-27
...........................\PE_AIC.m,2178,2011-03-27
...........................\Sym_Combin.m,380,2011-03-27
...........................\sym_representation.m,883,2011-03-27
...........................\variable_Combin.m,408,2011-03-27
...........................\variable_select.m,590,2011-03-27
...........................\X.xls,13824,2011-03-28
...........................\Y.xls,13824,2011-03-28
...........................\说明.txt,848,2011-03-28